Prescribed-Time Optimal Formation Control Using Fuzzy Reinforcement Learning for Second-Order Multiagent Systems.

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence
• Systems Engineering

Background:

• Multiagent systems require sophisticated control strategies for coordinated behavior.
• Existing prescribed-time control methods often face limitations with initial conditions.
• Optimal formation control is crucial for tasks involving coordinated movement and positioning.

Purpose of the Study:

• To investigate the prescribed-time (PT) optimal formation control for second-order multiagent systems.
• To develop a novel control scheme integrating reinforcement learning and fuzzy logic.
• To overcome limitations of existing PT control approaches, particularly regarding initial conditions.

Main Methods:

• A novel formation scheme combining reinforcement learning (RL) with a fuzzy logic system (FLS).
• Utilizing actor, critic, and identifier components within the RL-FLS framework.
• Introducing a prescribed performance function and filtered variable for error transformation.
• Developing an error transformation function for controller design independent of initial conditions.

Main Results:

• The proposed scheme ensures prescribed performance for the filtered error.
• All formation errors converge to a bounded region within the prescribed time.
• Satisfactory transient performance is achieved.
• The method demonstrates independence from initial tracking error and system dynamics.

Conclusions:

• The developed RL-FLS scheme effectively addresses the PT optimal formation control problem for second-order multiagent systems.
• The approach overcomes initial value limitations inherent in some PT control methods.
• The scheme guarantees prescribed performance and satisfactory transient behavior, validated through simulations.